KR102420597B1 - Autonomous driving system fail-safe utility and method thereof - Google Patents

Abstract

A fail-safe method of an autonomous driving system is disclosed. In this method, by using a first distribution map consisting of estimated positioning results projected from a single sensor input from a past time point to a current time point, and a current location location result currently input from the single sensor, the first distribution map for the single sensor performing fail-safe diagnostics; and when it is determined that the fail-safe diagnosis for the single sensor is successful or safe processing is possible, the most recently measured positioning results from each of the complex sensors including the single sensor based on the present time and performing fail-safe diagnosis on the composite sensors by analyzing a second distribution map composed of estimated positioning results projected to .

Description

Autonomous driving system with fail-safe function and method thereof

The present invention relates to an autonomous driving system having a fail-safe function and a method thereof.

Based on road map information built using GPS and various sensors (Radar, LiDAR, Camera, etc.) AVM, the autonomous driving system estimates positioning information that estimates the current location of a moving vehicle, and uses the estimated positioning information It is a system that controls the autonomous driving of a vehicle.

As positioning technologies for estimating the current location of a vehicle in such an autonomous driving system, there are satellite navigation, map matching, and the like. In these positioning techniques, the reliability of the estimated positioning information varies greatly depending on the surrounding environment of the vehicle.

The fluid reliability of such positioning information results in lowering the overall performance of the autonomous driving system.

As described above, the existing positioning technology estimates relatively accurate positioning information, but does not guarantee the reliability of the positioning information, so there is a possibility that erroneous positioning information may be utilized in the autonomous driving system.

Therefore, although a fail/safe function for estimating robust positioning information is essential to ensure reliability of positioning information, development of an autonomous driving system having a fail-safe function is still insufficient.

Accordingly, an object of the present invention is to provide an autonomous driving system and method having a fail-safe diagnosis function for estimating positioning information to ensure reliability of positioning information.

The problems to be solved in the present invention are not limited to those mentioned above, and other solutions not mentioned will be clearly understood by those of ordinary skill in the art from the following description.

In a fail-safe method of an autonomous driving system according to an aspect of the present invention for achieving the above object, estimated positioning results obtained by projecting the positioning results of the past time input from a single sensor to the current time point and the current input from the single sensor performing a fail-safe diagnosis on the single sensor using a first distribution map consisting of a positioning result of the current time point; and when it is determined that the fail-safe diagnosis for the single sensor is successful or safe processing is possible, the most recently measured positioning results from each of the complex sensors including the single sensor based on the present time and performing fail-safe diagnosis on the composite sensors by analyzing a second distribution map composed of estimated positioning results projected to .

An autonomous driving system having a fail-safe diagnosis function according to another aspect of the present invention diagnoses once whether a failure has occurred for the single sensor based on the frequency and distribution of positioning results input from a single sensor an initial diagnosis unit for performing an initial diagnosis; When it is determined that the initial diagnosis is successful, the estimated positioning results that project the positioning results of the past time input from the single sensor after the initial diagnosis to the current time point and the positioning result of the current time currently input from the single sensor a single sensor diagnosis unit for diagnosing fail-safe for the single sensor using the first distribution chart; and when it is determined that the fail-safe diagnosis for the single sensor is successful or safe processing is possible, the most recently measured positioning results from each of the complex sensors including the single sensor based on the present time and a complex sensor diagnosis unit for performing fail-safe diagnosis on the complex sensors by analyzing a second distribution map composed of estimated positioning results projected on the .

According to the present invention, by providing an autonomous driving system having a fail-safe diagnosis function for estimating positioning information to ensure reliability of positioning information, performance and reliability of the autonomous driving system can be improved.

Effects of the present invention are not limited to those mentioned above, and other effects not mentioned will be clearly understood by those of ordinary skill in the art from the following description.

1 is a block diagram of an unmanned autonomous driving system according to an embodiment of the present invention.
FIG. 2 is a diagram schematically illustrating an initial diagnosis process performed based on a distribution diagram of positioning results in the initial diagnosis unit shown in FIG. 1 .
FIG. 3 is a diagram schematically illustrating a process of performing fail-safe diagnosis on a single sensor by the single sensor fail-safe diagnosis unit shown in FIG. 1 .
FIG. 4 is a diagram schematically illustrating a process of performing fail-safe diagnosis on the complex sensor in the complex sensor fail-safe diagnosis unit illustrated in FIG. 1 .
5 is a flowchart illustrating a fail-safe diagnosis method for a positioning result in an autonomous driving system according to an embodiment of the present invention.

It should be noted that technical terms used herein are used only to describe specific embodiments, and are not intended to limit the present invention. In addition, the technical terms used in this specification should be interpreted in the meaning generally understood by those of ordinary skill in the art to which the present invention belongs, unless otherwise defined in this specification, and excessively inclusive. It should not be construed as a human meaning or in an excessively reduced meaning. In addition, when the technical terms used in this specification are incorrect technical terms that do not accurately express the spirit of the present invention, they should be understood by being replaced with technical terms that can be correctly understood by those skilled in the art. In addition, general terms used in the present invention should be interpreted as defined in advance or according to the context before and after, and should not be interpreted in an excessively reduced meaning.

Also, as used herein, the singular expression includes the plural expression unless the context clearly dictates otherwise. In the present application, terms such as "consisting of" or "comprising" should not be construed as necessarily including all of the various components or various steps described in the specification, some of which components or some steps are It should be construed that it may not include, or may further include additional components or steps. In addition, the suffixes "module", "unit" and "unit" for the components used in this specification are given or mixed in consideration of the ease of writing the specification. not.

The autonomous driving system of the present invention may perform sensor fusion positioning based on map matching.

In addition, the autonomous driving system of the present invention can improve the reliability of the system in the process of performing sensor fusion positioning and perform fail-safe diagnosis that enables robust and stable calculation (calculation or estimation) of positioning information.

In addition, since the fail-safe diagnosis of the present invention performs analytic redundancy-based fail-safe diagnosis (Analytic Redundancy-based Fault Diagnosis), additional hardware is not required.

In addition, the fail-safe diagnosis of the present invention includes an initialization diagnosis process, a single map matching-based fail-safe diagnosis process, and a complex map matching-based fail-safe diagnosis process, and includes detecting a fail sensor, identifying a fail sensor, and failing. Safe of the sensor is possible.

Accordingly, the reliability of systems that utilize positioning information such as autonomous driving can be greatly improved in the future.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of an unmanned autonomous driving system according to an embodiment of the present invention.

Referring to FIG. 1 , the autonomous driving system according to an embodiment of the present invention may include a fusion positioning module, a fail-safe diagnosis module 130 , and an output unit 150 .

The sensor fusion positioning module 110 may include a positioning sensor unit 112 and a positioning calculation unit 114 .

The positioning sensor unit 112 measures the current position of an unmanned autonomous vehicle (hereinafter, referred to as a 'vehicle'), and includes a plurality of positioning sensors of the same type or a plurality of positioning sensors of different types. can be composed of The positioning sensor may be, for example, a GPS sensor, a Radar sensor, a LiDAR sensor, a camera sensor, and the like, but is not limited thereto, and if it is a sensor related to the current location measurement of the vehicle, the type thereof is not limited.

The positioning calculation unit 114 fuses the position values measured by the plurality of sensors, and calculates a positioning result matching the fusion result with a road map prepared in advance. The positioning result may include vehicle location information on the road map, driving route information on the road map, and lane information based on the driving route information.

The fail-safe diagnosis module 130 performs fail-safe diagnosis on the positioning result input from the sensor fusion positioning module 110, and when a failure or abnormal symptom is diagnosed with respect to the positioning result, corresponding When a warning message is output and safety is diagnosed with respect to the positioning result, the positioning result can be corrected. This will be described in detail below.

The output unit 150 processes and outputs the warning message as visual, audible, or a combination thereof. In addition, the output unit 150 matches the normal positioning result passed according to the fail-safe diagnosis result performed to the fail-safe diagnosis module 130 or the positioning result restored according to the safe processing operation to the road map. The matching result is processed and outputted as visual, auditory, or a combination thereof. The output unit 150 may be a video output module, an audio output module, or a module combining them. Although not shown in the drawing, the video output module includes an image processing unit that converts the warning message or the corrected positioning result into image data such as text data and graphic data that can be output on an image screen, and a display unit that displays the image data (LCD, etc.), wherein the audio output module converts the warning message or the corrected positioning result into audio data such as voice data and the audio It may be configured to include a speaker unit that outputs data.

Hereinafter, the above-described fail-safe diagnosis module 130 will be described in detail with reference to FIGS. 2 to 4 .

The fail-safe diagnosis module 130 includes an initial diagnosis unit 132 and a fail-safe diagnosis unit 134 to perform fail-safe diagnosis on the positioning result input from the sensor fusion positioning module 110 . and a diagnosis result output unit 136 .

Initial diagnosis unit (132)

The initial diagnosis unit 132 may analyze the frequency and distribution of positioning results input from the sensor of the sensor fusion positioning module 110 to perform an initial diagnosis of the positioning result, that is, an initial fail-safe diagnosis.

The frequency may be defined as the number of times (number of measurements) of positioning results input from the positioning sensor during a preset time. Here, the preset time may be variously set according to a design, for example, may be 500 ms. That is, the initial diagnosis unit 132 counts the number of inputs of the positioning result during the preset time, and compares and analyzes the counted input number and the preset number (for example, 5 times) based on the result of analysis , it is diagnosed whether an initial failure has occurred for the positioning result. For example, if the number of inputs is 5 or more, the initial diagnosis for the positioning result may be determined as success, and if the number of inputs is less than 5 times, the initial diagnosis for the positioning result may be determined to fail.

When the initial diagnosis of the frequency of the positioning results is completed, the initial diagnosis unit 132 may perform an initial diagnosis on the positioning results input from the sensor fusion positioning module 110 based on the distribution of the positioning results. have.

Hereinafter, a process of performing an initial diagnosis based on the distribution of the positioning results will be described with reference to FIG. 2A .

,

,

)을 수집하고, 상기 수집된 과거 시점의 측위 결과들(

,

,

)이 현재 시점에서 투영(prediction)하여, 현재의 측위 결과들(이하, '추정 측위 결과들'이라 함.)을 추정한다. 여기서, 과거의 측위 결과들로부터 현재의 추정 측위 결과들을 추정하는 방법으로, 추측항법(Dead Reckoning: DR)이 이용될 수 있다.First, the initial diagnosis unit 132 determines the location results (

,

,

), and the location results of the collected past time points (

,

,

) is projected at the current time point to estimate the current positioning results (hereinafter referred to as 'estimated positioning results'). Here, as a method of estimating current estimated positioning results from past positioning results, dead reckoning (DR) may be used.

,

,

)이 추정되면, 상기 초기 진단 유닛(132)은 현재의 추정 측위 결과들(

,

,

)과 상기 센서 융합 측위 모듈에서 실제 입력되는 현재의 측위 결과(

)가 나타내는 위치들이 서로 떨어진 정도를 나타내는 분포(분포도 또는 분포 영역)(22 또는 24)의 표준편차(deviation)를 계산한다. Current estimated positioning results (

,

,

) is estimated, the initial diagnosis unit 132 determines the current estimated positioning results (

,

,

) and the current positioning result actually input from the sensor fusion positioning module (

Calculate the standard deviation of the distribution (distribution plot or distribution area) 22 or 24 indicating the degree to which the positions indicated by ) are separated from each other.

,

,

)과 현재의 측위 결과(

) 간들의 분포 범위가 좁을수록 초기 진단을 성공으로 판단할 확률이 높음을 알 수 있다.When the standard deviation is calculated, the initial diagnosis unit 132 compares the calculated standard deviation with a specific threshold value. If it exceeds, the initial diagnosis is judged as failing. To understand this schematically, FIG. 2(A) illustrates a case in which the initial diagnosis is determined to be a failure, and FIG. 2(B) illustrates a case in which the initial diagnosis is determined to be successful. That is, the current estimated positioning results (

,

,

) and the current positioning result (

), it can be seen that the narrower the distribution range between the livers, the higher the probability of judging the initial diagnosis as a success.

When it is determined that both the initial diagnosis of the frequency and distribution of the positioning results are successful, the initial diagnosis unit 132 requests the fail-safe diagnosis unit 134 to start fail-safe diagnosis. If at least one of the initial diagnosis for the frequency and distribution of the positioning results fails, the fail-safe diagnosis performed by the fail-safe diagnosis module 130 is not performed. Meanwhile, the initial diagnosis process by the initial diagnosis unit 132 may be performed only once for the first time.

Referring again to FIG. 1 , the fail-safe diagnosis unit 134 responds to a start request message (or a start request command) for the fail-safe diagnosis from the initial diagnosis unit 132 to fail-safe. Diagnosis can be performed.

In order to perform the fail-safe diagnosis, the fail-safe diagnosis unit 134 may include a single sensor fail-safe diagnosis unit 134A and a complex sensor fail-safe diagnosis unit 134B.

single sensor fail -safe diagnostic department (134A)

The single sensor fail-safe diagnosis unit 134A may perform fail-safe diagnosis of each single sensor. 3 schematically shows a process of performing fail-safe diagnosis for a single sensor in the single sensor fail-safe diagnostic unit shown in FIG. It is a case in which it is determined as a success, and (B) illustrates a case in which it is determined as a failure.

In order to perform fail-safe diagnosis of each single sensor, each single sensor may be assigned an identifier.

,

,

)을 수집하고, 수집된 과거의 측위 결과들(

,

,

)을 현재에서의 측위 결과들(

,

,

)(이하, 추정 측위 결과들이라 함)을 추정하고, 추정된 현재의 추정 측위 결과들(

,

,

)과 현재에서 실제 측정된 현재의 측위 결과(

)를 비교하여 페일 발생 여부를 진단할 수 있다. 이때, 페일을 진단하기 위한 임계값이 설정될 수 있으며, 상기 임계값은 현재의 측위 결과(

)를 기준으로 상기 대상 단일 센서의 스펙에 정의된 허용 오차 범위(30)(이하, 페일 허용 오차 범위 또는 제1 페일 허용 오차 범위)일 수 있다. The single-sensor fail-safe diagnosis unit 1354A uses the aforementioned dead reckoning (DR) to obtain past positioning results (

,

,

), and the collected past positioning results (

,

,

) to the current positioning results (

,

,

) (hereinafter referred to as estimated positioning results), and estimated current estimated positioning results (

,

,

) and the current positioning result (

) can be compared to diagnose whether a fail has occurred. At this time, a threshold value for diagnosing a failure may be set, and the threshold value is the current positioning result (

) based on the tolerance range 30 (hereinafter, a fail tolerance range or a first fail tolerance range) defined in the specification of the target single sensor.

,

,

)의 분포 영역(위치 분포도, 분포도 또는 제1 분포도)(

)의 전체가 상기 페일 허용 오차 범위(30) 내에 존재하면, 페일-세이프 진단을 성공으로 판단한다. 성공으로 판단되면, 상기 현재의 측위 결과(

)는 정상 측위 결과로 결정할 수 있다. The method of diagnosing whether a failure has occurred for the target single sensor is, for example, as shown in FIG. 3A, the current estimated positioning results (

,

,

) of the distribution area (location distribution, distribution or first distribution) (

) is within the fail tolerance range 30, it is determined that the fail-safe diagnosis is successful. If it is judged to be successful, the current positioning result (

) can be determined as a result of normal positioning.

,

,

)의 분포 영역의 전체가 상기 페일 허용 오차 범위(30)를 밖에 존재하면, 페일-세이프 진단을 페일로 진단한다. 만일 페일로 진단되는 경우, 세이프 처리동작을 수행할 수 있다. 여기서, 상기 세이프 처리동작은 페일로 진단된 측위 결과를 복구하는 처리 동작으로, 예를 들면, 현재의 측위 결과(

)는 폐기하고, 현재의 추정 측위 결과들(

,

,

) 중에서 어느 하나의 측위 결과를 선택하고, 선택된 측위결과를 상기 폐기된 현재의 측위 결과(

)를 복구한 측위 결과로 결정할 수 있다. 또는 현재의 추정 측위 결과들(

,

,

)이 각각 나타내는 위치들을 연결하는 도형을 생성하고, 생성된 도형의 중심 위치를 상기 폐기된 현재의 측위 결과(

)를 복구한 측위 결과로 결정할 수 있다.Conversely, as in (B) of Figure 3, the current estimated positioning results (

,

,

), the fail-safe diagnosis is diagnosed as fail when the entire distribution area of . If a fail diagnosis is made, a safe processing operation may be performed. Here, the safe processing operation is a processing operation of recovering the positioning result diagnosed as fail, for example, the current positioning result (

) is discarded, and the current estimated positioning results (

,

,

), select any one positioning result, and select the selected positioning result from the discarded current positioning result (

) can be determined by the restored positioning result. or current estimated positioning results (

,

,

) creates a figure connecting the positions respectively indicated, and sets the center position of the created figure to the discarded current positioning result (

) can be determined by the restored positioning result.

,

,

)로부터 상기 폐기된 현재의 측위 결과(

)를 복구한 측위 결과로 결정하는 것은 앞서 진행한 초기 진단 과정을 통과했기 때문이다. As such, the current estimated positioning results (

,

,

) from the discarded current positioning result (

) is determined as the restored positioning result because it has passed the initial diagnosis process.

,

,

)은 정상적인 측위 결과로 보아도 무방하므로, 이러한 추정 측위 결과들을 이용하여 복구 데이터를 결정하는 세이프 처리 동작을 수행할 있는 것이다.At least the current estimated positioning results (

,

,

) can be regarded as a normal positioning result, so a safe processing operation for determining recovery data can be performed using these estimated positioning results.

,

,

)의 분포 영역(

, 분포도)의 일부가 상기 페일 허용 오차 범위(30) 내에 존재하는 경우는, 완전한 페일로 진단할 수 없는 상황이다. 따라서, 이 경우도 현재의 측위 결과(

)를 정상적인 측위 결과로 판단할 수 있으며, 전술한 바와 같은 측위 결과를 복구하는 세이프 처리동작은 수행하지 않는다. 다만, 이상 징후 발생을 의심할 수 있는 경고 상황으로 진단할 수 있다.On the other hand, the current estimated positioning results (

,

,

) of the distribution area (

. Therefore, even in this case, the current positioning result (

) can be determined as a normal positioning result, and the safe processing operation for restoring the positioning result as described above is not performed. However, it can be diagnosed as a warning situation that can suspect the occurrence of abnormal symptoms.

As described above, when the fail-safe diagnosis is completed by the single-sensor fail-safe diagnosis unit 134A, the single-sensor fail-safe diagnosis unit 134A performs the positioning result or safe processing of each single sensor determined to be successful. The result of positioning restored by the operation may be transmitted to the composite sensor fail-safe diagnosis unit 134B and, at the same time, may request to start fail-safe diagnosis of the composite sensor.

the composite sensor fail -safe diagnostic department (134B)

The complex sensor fail-safe diagnosis unit 134B may perform fail-safe diagnosis on a complex sensor including different types of single sensors in response to a request from the single sensor fail-safe diagnosis unit 134A. have. 4 is a diagram schematically illustrating a process of performing fail-safe diagnosis on a composite sensor in the composite sensor fail-safe diagnostic unit shown in FIG. 1 , and FIG. 4A shows all single sensors included in the composite sensor. A case in which the fail-safe diagnosis is determined to be successful, (B) illustrates a case in which only some single sensors are determined to be successful among all the single sensors included in the complex sensor, and the remaining single sensors are determined to be failures.

The fail-safe diagnosis process for the composite sensor may be similar to the initial diagnosis process based on the above-described distribution diagram.

,

,

)을 수집하고, 수집된 가장 최근에 측정한 과거의 측위 결과들(

,

,

)을 추측 항법(DR)에 따라 현재에 투영하여, 현재의 추정 측정 결과들(

,

,

)을 추정한다.When it is assumed that the composite sensor includes the first to third sensors (sensor #1, sensor #2, and sensor #3), each sensor (sensor #1, sensor #2, and sensor #3) is based on the current The most recent past positioning results (

,

,

), and the most recently measured past location results (

,

,

) is projected to the present according to dead reckoning (DR), so that the current estimated measurement results (

,

,

) is estimated.

,

,

)의 분포도(

)(분포 영역, 위치 분포 또는 제2 분포도)의 표준 편차를 계산한 후, 표준 편차와 페일 허용 오차 범위(제2 페일 허용 오차 범위)를 나타내는 임계값을 비교하여, 표준 편차가 상기 임계값 이하인 경우(도 4의 (A)), 즉, 현재의 추정 측정 결과들(

,

,

) 모두가 페일 허용 오차 범위(40) 내에 존재하는 경우, 상기 현재의 추정 측정 결과들(

,

,

) 모두를 성공으로 판단하고, 표준 편차가 상기 임계값을 초과하는 경우, 상기 현재의 추정 측정 결과들(

,

,

) 모두를 페일로 진단한다.Then, the estimated current estimated measurement results (

,

,

) of the distribution (

) (distribution area, location distribution, or second distribution diagram) is calculated, and then the standard deviation is compared with a threshold value indicating the fail tolerance range (second fail tolerance range), and the standard deviation is less than or equal to the threshold value. case ((A) of FIG. 4), that is, the current estimated measurement results (

,

,

) are all within the fail tolerance range 40, the current estimated measurement results (

,

,

) are all judged as successful, and if the standard deviation exceeds the threshold, the current estimated measurement results (

,

,

) are diagnosed as failing.

)가 페일 허용 오차 범위(40)에 존재하는 경우, 페일로 진단하지 않고, 세이프 처리 동작을 수행하여, 상기 추정 측위 결과(

)를 페일로 판단되는 나머지 추정 측위 결과들(

,

)을 복구한 추정 측위 결과(

)로 결정한다. 이때, 페일 허용 오차 범위(40)에 존재하는 추정 측위 결과가 복수인 경우, 페일 허용 오차 범위(40)에 존재하는 복수의 추정 측위 결과를 생성한 각각 센서들 중 신뢰도가 가장 높은 센서로부터 추정된 추정 측위 결과를 복구된 측위 결과로 결정하는 세이프 처리 동작을 수행할 수 있다. 여기서, 신뢰도가 가장 높은 센서는 각 센서의 스펙에 존재하는 허용 오차 범위가 가장 작은 센서로 정의할 수 있다. 한편, 상기 페일 허용 오차 범위에 존재하는 센서들을 식별하기 위해 다양한 기하학적 방법이 활용될 수 있다. On the other hand, as shown in FIG. 4B , the positioning measured by some sensors (hereinafter, the first sensor (sensor #1)) among the sensors (sensor #1, sensor #2, and sensor #3) included in the complex sensor. The estimated positioning result (

) is within the fail tolerance range 40, the estimated positioning result (

) of the remaining estimated positioning results judged to be fail (

,

), the estimated positioning result (

) to be determined. At this time, when there are a plurality of estimated positioning results existing in the fail tolerance range 40 , the estimation is performed from the sensor with the highest reliability among the sensors that have generated a plurality of estimated positioning results existing in the fail tolerance range 40 . A safe processing operation of determining the estimated positioning result as the restored positioning result may be performed. Here, the sensor with the highest reliability may be defined as the sensor having the smallest allowable error range existing in the specification of each sensor. Meanwhile, various geometrical methods may be used to identify sensors existing within the fail tolerance range.

Diagnostic result output unit (136)

Referring back to FIG. 1 , when all fail-safe diagnoses for single sensors and complex sensors are completed by the fail-safe diagnosis unit 134 , the diagnosis result output unit 136 displays the diagnosis results in a plurality of fail-safe diagnoses. A message corresponding to the diagnosis result may be provided to the user with reference to the table below classified by level.

fail level
(Fail Level) Condition Countermeasure availability
(Availability) Level0 steady state normal output Available Level1 anomalies occur alert Available Level2 Some sensor fail Safe handling Available Level3 Total sensor fail fail handling can not be used

Table 1 above is a table defining a fail level according to a fail-safe diagnosis result according to an embodiment of the present invention.

)과 페일 오차 범위(30)가 일부 겹치는 경우이다. Level2은 페일-세이프 진단에서 세이프 처리가 가능한 페일이 발생한 경우이다. Level2은 세이프 처리가 불가능한 Fail이 발생한 경우이다. Level 0, 1, 2는 Fail이 발생한 경우에는 측위 결과의 사용이 가능하지만 Level 3은 측위결과의 사용이 불가능하며 시스템은 다시 초기화 진단을 진행하게 된다. In one example, the fail level may consist of four steps. Level 0 is when the positioning result is normal. That is, it is a case in which all of the initial diagnosis and the fail-safe diagnosis of the single sensor and the complex sensor are determined to be successful. Level 1 is a case in which anomaly warning is indicated. That is, in the fail-safe diagnosis for a single sensor, the distribution area (

) and the fail error range 30 partially overlap. Level 2 is a case where a fail that can be safely processed occurs in the fail-safe diagnosis. Level 2 is a case where a Fail that cannot be safely processed occurs. In Level 0, 1, and 2, if a Fail occurs, positioning results can be used, but in Level 3, positioning results cannot be used, and the system proceeds with initialization diagnosis again.

5 is a flowchart illustrating a fail-safe diagnosis method for a positioning result in an autonomous driving system according to an embodiment of the present invention. In describing each of the following steps, the contents described with reference to FIGS. 1 to 4 The overlapping contents will be briefly described or omitted.

Referring to FIG. 5 , first, in step S510 , the frequency and distribution of positioning results input from a single sensor included in the sensor fusion positioning module 110 are analyzed to perform an initial diagnosis on the positioning results. This initial diagnosis may be performed once for each sensor.

As described above, the frequency may be the number of times (the number of measurements) of positioning results input from the positioning sensor during a preset time. In this case, for example, if the number of inputs is 5 or more, the initial diagnosis for the positioning result may be determined as success, and if the number of inputs is less than 5 times, the initial diagnosis for the positioning result may be determined to fail. .

,

,

)을 현재 시점에서 투영(prediction)한 추정 측위 결과들과 상기 센서에서 실제 입력되는 현재의 측위 결과(

)가 나타내는 위치들이 서로 떨어진 정도를 나타내는 분포(분포도 또는 분포 영역)(22 또는 24)의 표준편차(deviation)를 계산하고, 계산된 표준편차와 특정 임계값을 비교한다.When the initial diagnosis of the frequency is completed, the initial diagnosis may be performed based on the distribution of the positioning results. Specifically, the positioning results of the past time (

,

,

) projected (prediction) from the current time point and the current positioning result actually input from the sensor (

) calculates the standard deviation of the distribution (distribution chart or distribution area) 22 or 24 indicating the degree to which the positions are separated from each other, and compares the calculated standard deviation with a specific threshold.

When the deviation is less than or equal to a specific threshold, the initial diagnosis is determined as successful, and when the deviation exceeds the specified threshold, the initial diagnosis is determined as fail.

If the initial diagnosis based on the frequency and distribution of the positioning results is successful, in step S520, the estimated positioning results obtained by projecting the positioning results of the past time input from a single sensor after the initial diagnosis to the current time point and the single A fail-safe diagnosis is performed for the single sensor by using a first distribution map composed of the positioning result of the current time input currently from the sensor.

Specifically, based on a result of comparative analysis of a threshold value indicating a first fail tolerance range defined for the single sensor based on the standard deviation of the first distribution and the current positioning result, the fail for the single sensor -Perform safe diagnosis. For example, if the standard deviation exceeds the threshold, the fail-safe diagnosis for the single sensor is determined as fail, and if the standard deviation is below the threshold, the fail-safe diagnosis for the single sensor is performed. It is judged as a case of success or safe processing.

If it is determined that the fail-safe diagnosis for the single sensor is successful or safe processing is possible, in step S530, the most recently measured positioning based on the present from each of the multiple sensors including the single sensor A fail-safe diagnosis is performed on the composite sensors by analyzing a second distribution map composed of estimated positioning results projected on the current time point.

Specifically, the fail-safe diagnosis of the composite sensor is performed based on a result of comparing and analyzing the standard deviation of the second distribution and a threshold value indicating a predefined second fail tolerance range. For example, if the standard deviation exceeds the threshold, fail-safe diagnostics for all of the composite sensors are determined as fail, and if the standard deviation is less than or equal to the threshold, fail for all of the composite sensors. - Assess the Safe Diagnosis as a success.

Then, when fail-safe diagnosis of all of the composite sensors is completed, in step S540, the diagnosis result is output.

The diagnosis result may be output in the form of a message composed of, for example, four fail levels (Level0 to Level3).

)과 페일 오차 범위(30)가 일부 겹치는 경우이다. Level2은 페일-세이프 진단에서 세이프 처리가 가능한 페일이 발생한 경우이다. Level2은 세이프 처리가 불가능한 Fail이 발생한 경우이다. Level 0, 1, 2는 Fail이 발생한 경우에는 측위 결과의 사용이 가능하지만 Level 3은 측위결과의 사용이 불가능하며 시스템은 다시 초기화 진단을 진행하게 된다. Level 0 is when the positioning result is normal. That is, it is a case in which all of the initial diagnosis and the fail-safe diagnosis of the single sensor and the complex sensor are determined to be successful. Level 1 is a case in which anomaly warning is indicated. That is, in the fail-safe diagnosis for a single sensor, the distribution area (

) and the fail error range 30 partially overlap. Level 2 is a case where a fail that can be safely processed occurs in the fail-safe diagnosis. Level 2 is a case where a Fail that cannot be safely processed occurs. In Level 0, 1, and 2, if a Fail occurs, positioning results can be used, but in Level 3, positioning results cannot be used, and the system proceeds with initialization diagnosis again.

Meanwhile, it should be understood that the block diagram of FIG. 1 showing an autonomous driving system having a fail-safe function embodies the principle of the invention from a functional point of view. Similarly, the flow diagram of FIG. 5 is to be understood as representing various processes that may be tangibly embodied on a computer-readable medium and performed by a computer or processor, whether or not a computer or processor is explicitly shown.

Blocks of FIG. 1 indicated by a processor or a similar concept may be provided by use of hardware capable of executing software as well as dedicated hardware.

When the blocks of FIG. 1 are implemented by a processor, the functions of the blocks shown in FIG. 1 may be provided by a single dedicated processor, a single shared processor, or a plurality of individual processors, some of which may be shared.

In addition, the clear use of terms presented as processor, control, or similar concepts should not be construed as exclusively referring to hardware capable of executing software, and without limitation, digital signal processor (DSP) hardware, ROM for storing software. It should be understood to implicitly include (ROM), RAM (RAM) and non-volatile memory. Of course, other hardware for general use may also be included.

The above description is merely illustrative of the technical idea of the present invention, and various modifications, changes and substitutions are possible within the scope that does not depart from the essential characteristics of the present invention by those of ordinary skill in the art to which the present invention pertains. will be. Accordingly, the embodiments disclosed in the present invention and the accompanying drawings are for explanation rather than limiting the technical spirit of the present invention, and the scope of the technical spirit of the present invention is not limited by these embodiments and the accompanying drawings. . The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

Claims (12)

Fail-safe ( fail-safe) diagnostics; and
When it is determined that the fail-safe diagnosis for the single sensor is successful or safe processing is possible, the most recently measured positioning results based on the present from each of the complex sensors including the single sensor are displayed at the current time. performing fail-safe diagnosis on the composite sensors by analyzing a second distribution map composed of the projected estimated positioning results
A fail-safe method of an autonomous driving system comprising a.
The method of claim 1, wherein the performing fail-safe diagnosis for the single sensor comprises:
Fail-safe diagnosis for the single sensor based on a result of comparing and analyzing the standard deviation of the first distribution map and a threshold value indicating the first fail tolerance range defined for the single sensor based on the current positioning result A fail-safe method of an autonomous driving system, characterized in that performing the
The method of claim 2, wherein the performing fail-safe diagnosis for the single sensor comprises:
When the standard deviation exceeds the threshold value, the fail-safe diagnosis for the single sensor is determined as fail, and if the standard deviation is less than or equal to the threshold value, the fail-safe diagnosis for the single sensor is determined as success Fail-safe method of an autonomous driving system, characterized in that the step.
The method of claim 1, wherein the performing fail-safe diagnosis for the single sensor comprises:
When it is determined that the fail-safe diagnosis for the single sensor is fail, the positioning result of the current time is discarded, a figure connecting the positions respectively indicated by the estimated positioning results is generated, and the center position of the generated figure is set as the A fail-safe method of an autonomous driving system, comprising the step of performing a safe process for determining the discarded positioning result of the current time point as the restored positioning result.
The method of claim 1 , wherein the performing fail-safe diagnosis on the composite sensors comprises:
Autonomous driving, characterized in that the step of performing fail-safe diagnosis on the composite sensor based on a result of comparing and analyzing the standard deviation of the second distribution and a threshold value indicating a predefined second fail tolerance range The fail-safe method of the system.
The method of claim 5, wherein the performing fail-safe diagnosis of the composite sensors comprises:
If the standard deviation exceeds the threshold, fail-safe diagnosis for all of the composite sensors is determined as fail, and if the standard deviation is less than or equal to the threshold, fail-safe diagnosis for all of the composite sensors is performed. A fail-safe method of an autonomous driving system, characterized in that it is a step of determining success.
The method of claim 5, wherein the performing fail-safe diagnosis of the composite sensors comprises:
Estimated positioning results estimated from some of the sensors of the composite sensors exist within the second fail tolerance range, so it is determined as a success, and estimated positioning results estimated from the remaining sensors exist outside the second fail tolerance range. , when it is determined as a failure, performing a safe process for determining the estimated positioning result estimated by the sensor with the highest reliability as the restored positioning result from the estimated positioning results determined to be failure. 's fail-safe method.
The method of claim 1 , wherein before performing fail-safe diagnostics for the single sensor,
The fail-safe method of the autonomous driving system, further comprising: performing an initial diagnosis of diagnosing once whether or not a failure has occurred in the single sensor based on the frequency and distribution of positioning results input from the single sensor .
The method of claim 8, wherein in the performing the initial diagnosis, when it is determined that the initial diagnosis is a fail,
The fail-safe method of an autonomous driving system, characterized in that the fail-safe diagnosis of the single sensor and the fail-safe diagnosis of the complex sensors are not performed.
The method of claim 8, wherein the frequency of the positioning results input from the single sensor,
The fail-safe method of an autonomous driving system, characterized in that it is the number of times the single sensor measures the positioning results during a preset time.
The method of claim 8, wherein the distribution of the positioning results input from the single sensor,
A fail-safe method of an autonomous driving system, characterized in that it is a distribution map comprising estimated positioning results obtained by projecting the positioning results of the past time input from a single sensor to the current time point and the current positioning results currently input from the single sensor.
an initial diagnosis unit configured to perform an initial diagnosis of diagnosing whether or not a failure has occurred with respect to the single sensor once based on the frequency and distribution of positioning results input from a single sensor;
When it is determined that the initial diagnosis is successful, the estimated positioning results that project the positioning results of the past time input from the single sensor after the initial diagnosis to the current time point and the positioning result of the current time currently input from the single sensor a single sensor diagnosis unit for diagnosing fail-safe for the single sensor using the first distribution chart; and
When it is determined that the fail-safe diagnosis for the single sensor is successful or safe processing is possible, the most recently measured positioning results based on the present from each of the complex sensors including the single sensor are displayed at the current time. A complex sensor diagnosis unit that analyzes a second distribution map composed of the projected estimated positioning results and performs fail-safe diagnosis on the complex sensors
Fail-safe autonomous driving system with a function comprising a.

Images